People + Strategy Journal

Winter 2022

Data-Driven Approaches to Diversity, Equity and Inclusion

Many leaders struggle to use a data-driven approach to DE&I. Choosing not to collect data, reporting only favorable trends or weighting quantitative over qualitative data are common errors.

By ​Laura Morgan Roberts and Melissa Thomas-Hunt
Data-Driven Approaches to Diversity, Equity and Inclusion

​Laura Morgan Roberts, Ph.D., and Melissa Thomas-Hunt, Ph.D., are colleagues at the University of Virginia Darden School of Business, with over two decades of experience as researchers and practitioners in the fields of diversity, equity and inclusion (DE&I). Thomas-Hunt has focused extensively on leading teams, negotiations, status and expertise. Morgan Roberts has focused on identity, authenticity, race and inclusion. Both have engaged in practice as well—Thomas-Hunt as Chief Diversity Officer for Airbnb, Vice Provost of Vanderbilt University and Chief Diversity Officer of the Darden School; and Morgan Roberts as a consultant. 

What are best practices in designing data-driven DE&I initiatives that will increase representation, advancement, engagement and equity? Here's what we've observed about challenges in how leaders deal with DE&I data. Most avoid the data altogether, often choosing to not even collect it or they parse and report only favorable trends without peeling back layers of the onion that expose disparities, inequality and pain. There's also a tendency to weight large quantitative datasets over qualitative data, sometimes discounting a voluminous accumulation of anecdotes about lived experiences, which can provide compelling insight regarding context and action. To improve the quality and impact of DE&I data, we recommend partnering with social scientists who are skilled in examining DE&I dynamics and relevant outcomes, guided by a spirit of inquiry and organizational transparency, and committed to protecting the vulnerability of those numerical minorities and/or others with less positive experiences at work. 

Here we highlight seven data-driven key practices that strengthen the focus and impact of leaders' DE&I efforts. 

First, you must be willing to ask the question. Many organizations lack DE&I focus and ambition because their leaders are unwilling to ask the questions that data could help to answer. Fear and resistance to change push leaders to avoid gathering data about patterns of representation, progress, belonging and overall employee engagement. Concerns about having to address or fix what is found paralyze efforts to understand the nuances of individuals' disparate experiences. You can't address what you don't identify as a challenge or issue, and robust data on diversity, inclusion and equity help to illuminate pressing problems in talent management and social impact. For example, one of us recently spoke to the CEO of a relatively small firm that is scaling who whispered, "I don't actually know who's here! Do you think we should do a survey of demographics or identities?" A survey is a starting point, and we encourage the usage of surveys to gather relevant data. You can also invest in human capital management software that allows you to systematically capture your employees' identities and map that against the other data you have about them. Your approach depends how repeatable and embedded you want this data-driven DE&I effort to be in the fabric of your organization. You can do this even if you are relatively small and scaling. 

Second, design your data collection efforts with strategic intent, and acknowledge the choices you are making at the onset. Contrary to the popular statement, the numbers do not speak for themselves. Data are never neutral. The people collecting the data have a point of view, and their perspectives and agenda shape every aspect of data collection, analysis, interpretation and reporting. As a leader, what questions about DE&I is your organization asking? What response categories have you offered for people to identify themselves and share their experiences, and why? When, where and from whom is your organization proactively soliciting systematic input? Are you carefully considering how to reach out to encourage members of groups with small numerical representation? Are you signaling that their input matters and will be valued? Are you artificially and unnecessarily constraining the answer set? How does your organization account for a relatively smaller pool of respondents among underrepresented minorities? These design choices will impact what you learn, from whom, and the narrative that is crafted from your data. 

Many get stuck here on what you can and can't ask in your data collection. It depends where you are in the world. However, almost anything is possible if you are motivated to understand the broad array of employee experiences across the organization. In not finding ways to collect the data broadly, some individuals' experiences get rendered invisible. Are you selectively telling some stories and not others? To be clear, there are data collecting and reporting challenges that exist, but there are ways to work through them and begin to make progress with broadening or deepening your understanding. However, people don't even understand what are the true obstacles because they haven't started down the path. 

Third, disaggregate your data. Increasingly organizations are choosing to openly share employee engagement and belonging data. However, the data is shown across the company or perhaps by department. Rarely is the data cut by gender, race/ethnicity or age. The decision not to disaggregate the data is often made unwittingly. Many organizational leaders can't imagine that there would be differences in the experience of individuals from different groups. This in itself can feel like an identity abrasion because some individuals feel like their experiences are obscured and rendered unworthy of note. To really understand the landscape of experiences, you need to cut the data by subcategories. Some organizations are even beginning to look at sexual orientation, gender identity, parent/caregiver status and disability. 

Fourth, use multiple forms of data to inform your assessment, especially when seeking insight about marginalized workers and impacted communities. Avoid over-relying on quantitative data—when it comes to DE&I, not everything that counts can be counted, and small numbers are often dismissed due to an inability to conduct statistically significant comparisons. Qualitative data are uniquely important for providing contextualized accounts of people's daily experiences. They are especially helpful with interpreting patterns of responses among a broader group, and exploring how people who are extremely satisfied or successful differ from those who are extremely dissatisfied or struggling. Using interviews, focus groups, case studies and other forms of qualitative inquiry can illuminate how issues of underrepresentation and exclusion have emerged in the organization over time, why they persist, and when and where they manifest. Yet we have noted that many leaders actively discount people's accounts of exclusion, dismissing them as a singular experience that was likely misinterpreted or blown out of proportion. This not only invalidates the person who is already feeling marginalized, it hampers organizational learning essential for best practices in DE&I. People will refrain from speaking out about offensive and hurtful comments, ongoing disparities and adverse social impact if they feel their stories are not believed or used to catalyze change. Leaders who dismiss single stories overlook systemic patterns in toxic behavior by individuals on one extreme, or subtle processes that reinforce inequality on the other hand. Often dismissed as anecdotes, people's narratives offer a valuable lens into organizational culture, climate and experiences of inclusion and exclusion. They can allow people to see themselves in the data and inspire action. 

Fifth, don't stockpile data for data's sake; instead, use your DE&I data to inform action. The insatiable desire for more and more data, analysis paralysis and prove-it-again tendencies stifle deep change. While big data is impressive in size and scale, programmatic interventions should also be tracked on a local level and at an incremental pace. If the data begins pointing in a direction, move forward. Place a small bet. Stand up an initiative or provide a pilot workshop. You can always refine as the data becomes more concrete. 

Sixth, data should be used to drive strategic planning, but should also be considered part of a learning orientation of discovery, documentation and interventions. Engage in partnerships with social scientists to who can help you to develop evidence-based practices that truly move the needle. Leading edge academic experts are eager to engage. It's easier to garner their assistance than one might think. Not only have they examined patterns outside of your organization, their arms-length relationship as trusted learning partners will help leaders to gather data with rigor and relevance, ask probing questions that challenge organizational assumptions, and bring out the perspectives of people who may be less vocal in conversations with leaders and coworkers. For example, we have collaborated with organizations in conducting focus groups, gathering survey data, designing field experiments and assessing interventions.

Finally, results should be made publicly available. Gathering and burying one's own numbers, out of fear that they don't reflect favorably, will serve only to derail efforts to increase representation, build inclusion, ensure equity and uphold justice at work. It's okay to show that there is work to be done. Goodwill comes from transparency. Remember, individuals are living the state of affairs. They already know what's working and what's not. Showing the data does not create the friction points. The data doesn't create the problem. It also doesn't fix the challenging areas. It does create trust from transparency and fosters a belief that the organization is ready to move forward with action.